summaryrefslogtreecommitdiff
path: root/src (follow)
Commit message (Collapse)AuthorAgeFilesLines
* Crank libcrypto major since my previous commit changed the size of thejsing2014-06-242-2/+2
| | | | ChaCha context. Other changes will also ride this crank.
* If a chacha operation does not consume all of the generated key stream,jsing2014-06-246-14/+92
| | | | | | | | | | | | | | | | ensure that we save it and consume it on subsequent writes. Otherwise we end up discarding part of the key stream and instead generate a new block at the start of the next write. This was only an issue for callers that did multiple writes that are not multiples of 64 bytes - in particular, the ChaCha20Poly1305 usage does not hit this problem since it performs encryption in a single-shot. For the same reason, this is also a non-issue when openssl(1) is used to encrypt with ChaCha. Issue identified by insane coder; reported to bugs@ by Joseph M. Schwartz. ok beck@
* Extend the chacha regress to cover the ChaCha interface, in addition to thejsing2014-06-241-22/+99
| | | | | | | single-shot CRYPTO_chacha_20() interface (the ChaCha interface was already tested via the EVP regress, but not extensively). The additional ChaCha tests include single-shot writes, along with partial/single-byte writes that currently fail due to a bug in the underlying implementation.
* Some KNF.jsing2014-06-242-12/+24
|
* Replace 48 lines of code with a single inet_pton() call. The previousjsing2014-06-242-96/+6
| | | | | | | | | | | | | handrolled version could not even make use of sscanf(), since that would not work with a certain antiquated compiler. It is worth noting that there is a tiny change in behaviour - previously calling BIO_get_host_ip() with something that looked like it might be a valid IP address (for example, "1." or even ".") would result in it returning failure rather than trying a BIO_gethostbyname() - now we'll always try a BIO_gethostbyname() if it was not a valid IPv4 address. ok beck@ miod@ deraadt@
* Actually make BIO_set_tcp_ndelay() work - TCP_NODELAY will not magicallyjsing2014-06-242-32/+6
| | | | | | appear by itself. ok beck@ miod@
* Fix memory leak.logan2014-06-241-2/+4
| | | | | | Thanks to Brenk Cook. OK from miod@
* Since this is a library, place issetugid() before every getenv()deraadt2014-06-2310-26/+48
| | | | ok miod
* unbreak build of getentropy_sysctl - we need linux/sysctl.h, andbeck2014-06-232-36/+42
| | | | RANDOM_UUID is an enum member.
* unbreak - main needs to be extern in here somewhere.beck2014-06-232-2/+4
|
* KNF, particularly wrapped lines of calls to PEM_read_bio_FOO() andguenther2014-06-222-48/+88
| | | | | | multiline comments ok jsing@
* Add regress tests for BIO_get_host_ip().jsing2014-06-221-2/+70
|
* BIO_sock_init() no longer does anything, so stop calling it.jsing2014-06-222-20/+2
|
* Just use SOMAXCONN and IPPROTO_TCP, since we know we have them.jsing2014-06-222-28/+8
|
* In BIO_get_port(), use strol() with appropriate range checks rather thanjsing2014-06-226-68/+100
| | | | | | | an atoi() followed by an unsigned short cast. This stops things like "-1" and "66536" from being considered to be "valid" port numbers. ok beck@ deraadt@
* Add a skeleton regress for crypto/bio, which currently only coversjsing2014-06-223-1/+94
| | | | | BIO_get_port() and fails since the current code believes that "-1" is a valid port.
* Hook in the aead regress.jsing2014-06-221-1/+2
|
* nuke unused test programs; ok jsingderaadt2014-06-224-92/+4
|
* More KNF.jsing2014-06-222-10/+10
|
* KNF.jsing2014-06-226-416/+406
|
* KNF.jsing2014-06-2212-1922/+2162
|
* More KNF.jsing2014-06-226-18/+18
|
* matthew reminds me to update regress to reflect current spectedu2014-06-211-10/+3
|
* repair indentation for an inner loop; shorten some macros and variablederaadt2014-06-212-258/+270
| | | | | names to shorten line lengths ok beck
* always compare memcmp against 0, for clarity.tedu2014-06-216-14/+14
|
* loosen the spec for timingsafe functions slightly, so as to nottedu2014-06-211-10/+5
| | | | artificially constrain alternative implementations. ok deraadt
* Pull the code that builds a DTLS sequence number out into its own functionjsing2014-06-216-46/+58
| | | | | | to avoid duplication. Also use fewer magic numbers. ok miod@
* Specify the correct strength bits for 3DES cipher suites.jsing2014-06-212-30/+26
| | | | | | From OpenSSL. ok miod@
* Add DTLS support to ssltest and wire up some regress tests.jsing2014-06-212-8/+41
| | | | ok miod@
* Switch to the ISC licensed versions of these files, which Google has madejsing2014-06-214-202/+52
| | | | | | available via boringssl. ok deraadt@
* Pull out the sequence number selection and handle this up front. Also, thejsing2014-06-212-18/+12
| | | | correct record is already known, so avoid reassignment.
* More KNF and clean up.jsing2014-06-212-26/+18
|
* More KNF.jsing2014-06-2116-122/+112
|
* KNFmiod2014-06-216-248/+274
|
* KNFmiod2014-06-216-356/+372
|
* Fix memory leak in error path.logan2014-06-212-4/+4
| | | | OK from miod@
* Protect explicit_bzero() from link-time optimizationmatthew2014-06-211-7/+10
| | | | | | | | | | | | | | | | | | | | Modern compiler toolchains are capable of optimizing even across translation unit boundaries, so simply moving the memory clearing into a separate function is not guaranteed to clear memory. To avoid this, we take advantage of ELF weak symbol semantics, and insert a call to an empty, weakly named function. The semantics of calling this function aren't determinable until load time, so the compiler and linker need to keep the memset() call. There are still ways a toolchain might defeat this trick (e.g., optimistically expecting the weak symbol to not be overloaded, and only calling memset() if it is; promoting weak symbols to strong symbols at link-time when emitting a static binary because they won't be interposed; implementing load-time optimizations). But at least for the foreseeable future, these seem unlikely. ok deraadt
* hash in correct pointerderaadt2014-06-212-4/+4
|
* Remove the OPENSSL_*cap getenv's. A program should not be able toderaadt2014-06-206-50/+10
| | | | | change the behaviour of the library in such a complicated fashion. ok miod
* wrap getenv OPENSSL_ALLOW_PROXY_CERTS in an issetugid check, to protectderaadt2014-06-202-4/+4
| | | | | setuid applications from being fooled. ok miod
* KNFbeck2014-06-202-96/+112
|
* indentderaadt2014-06-202-4/+4
|
* rearrange so that the main function with the important comments is at the topotto2014-06-202-156/+160
| | | | ok deraadt@ beck@
* Work in progress on how to deal with the inherit unreliability ofbeck2014-06-202-0/+878
| | | | | | /dev/urandom. Does well in the fallback case. Get it in tree so it can be worked on. ok otto@ deraadt@
* Remove OPENSSL_instrument_halt and OPENSSL_far_spin, which both mightmiod2014-06-202-140/+0
| | | | have been used under DJGPP in the previous century (if at all).
* Fix incorrect bounds check in amd64 assembly version of bn_mul_mont();miod2014-06-204-8/+8
| | | | | noticed and fix by Fedor Indutny of Joyent ( https://github.com/joyent/node/issues/7704 )
* Simple regress test for the amd64 bn_mul_mont bug found by Joyentmiod2014-06-203-2/+86
| | | | | ( https://github.com/joyent/node/issues/7704 ), about to be fixed in libcrypto.
* Move the crypto/bn regression test one directory deeper in preparation formiod2014-06-203-7/+15
| | | | it getting siblings.
* arc4random: hard fail with raise(SIGKILL) if getentropy() returns -1matthew2014-06-201-3/+4
| | | | | | | Allow other non-zero return values in case we change our mind to return an ssize_t byte count instead of simple success/fail. ok deraadt, djm
* convert CRYPTO_memcmp to timingsafe_memcmp based on current policy favoringtedu2014-06-1918-44/+44
| | | | | | libc interfaces over libcrypto interfaces. for now we also prefer timingsafe_memcmp over timingsafe_bcmp, even when the latter is acceptable. ok beck deraadt matthew miod
generated by cgit v1.2.3-55-g6feb (git 2.39.1) at 2025-04-12 20:12:48 +0000